From: SCN User (mentifex@scn.org)
Date: Mon Oct 01 2001 - 13:54:18 MDT
>> What architecture is Mind based on? Agents?
>
ESY (reformatted by ATM for readability):
> Well, I've never been able to figure it out
> from looking at Mentifex's diagram at
> http://mind.sourceforge.net/
> and I still can't figure it out from looking at
> http://cvs.sf.net/cgi-bin/viewcvs.cgi/mind/Modules/Mind.f?rev=1.1.1.1
> except that it seems to involve classical-AI processing
> of Chomskian grammars by assignment of individual
> computational tokens to words, which
> is of course one of the all-time GOFAI mistakes.
ATM:
Good Old-Fashioned Artificial Intelligence (GOFAI)
may have failed with attempts to "tokenize" words, but
http://sourceforge.net/projects/mind is a radically new
approach which _conceptualizes_ words (and, for now, only words)
while leaving associative hooks available for fleshing out
the word-concepts with additional and primitive sensory data.
This preparation for an enhanced sensorium is why the AI Mind
is written not only in JavaScript for easy Web distribution,
but also in Win32Forth for implementation in robots (which
I do not believe has actually happened yet).
ESY:
> I don't think this AI project has gotten up to
> the agent level yet, and if there's any contamination
> by GISAI or modern cognitive science, it is not evident.
ATM:
GISAI is not familiar to me. As for cognitive science, the
AI Mind is a totally new approach but rooted in neuroscience.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT